@aaronventure just the amount applied needs to be "polyphonic", not the type of modulation itself
but I could do that with an "intensity" pma node inside the oscillator network where I already have the velocity information
I also have a hardcoded LFO in the faust oscillator but it's not time synced among all notes of course. that might be a benefit considering playability and sound design capability
I guess the CPU overhead of running seperate LFO oscillators on each voice is superficial